By: Ian McWilliams, PhD
Melting ice sheet may expose cold war base, hazardous waste
During the Cold War, the US Army Corps began a top-secret mission to determine the capability of launching nuclear missiles at Russia from a base in Greenland. The military base constructed for this mission, named Camp Century, lies approximately 125 miles inland from the Greenland coast and was later abandoned in 1964 after the Joint Chiefs of Staff rejected the plans to create a nuclear base. When soldiers abandoned the base, it was thought that leftover fuel and waste material would be safely interred, buried under ice for thousands of years.
However, climate change has now threatened those plans. The increased ice melt could reveal the base as early as 2090 and it is estimated that tens of thousands of gallons of diesel fuel, wastewater, sewage, and other chemicals could be exposed. Adding to concerns is the nuclear generator housed in the frozen base. Although the base never became a site for nuclear weapons, the low-level radioactive coolant from the nuclear generator is still stored in the base. If ice melt continues to occur at an accelerated rate, some have expressed concern that these chemicals could be released into the environment by seeping into waterways causing a potential environmental catastrophe. (Stephen Feller, UPI)
Mouse microbe may make scientific studies harder to replicate
Reproducibility is an issue that has been the subject of much debate in the scientific community recently. Now, scientists are concerned that the microbiome may further complicate the issue. The collection of commensal microorganisms that reside on or within the body is referred to as microbiota, and it is now well known to affect the health of the host. Although researchers have taken meticulous steps to ensure that experimental animals are housed in identical conditions, including sterile bedding, strict temperature control, and standard light cycles, determining experimental variability due to differences in their microbiome have remained elusive. As researchers explore the issue further they have found that mice from different vendors have very different compositions of bacteria in their gut that could explain some inconsistencies in researchers’ experiments.
Although it is not mandated, taking steps to control for microbiome may aid in the reproducibility crisis. Segmented filamentous bacteria (SFB) have been identified as a notable concern, and some vendors are providing SFB positive or SFB negative animals separately. Although it is unlikely that SFB is the only culprit for differences in studies, researchers continue to explore new variables in rodent husbandry in an effort to improve reproducibility of scientific results. To add to the dilemma, because the species that constitute the microbiome are constantly changing, it is difficult to characterize, and impossible to standardize. Since mice share their microbes through eating each other’s feces, cage-mates can have similar microbiomes that provide natural microbiota normalization for littermates. (Kelly Servick, Science)
Spiking genomic databases with misinformation could protect patient privacy
New initiatives, like the Precision Medicine Initiative (PMI), are helping to cultivate the human genome into usable sets of data for research purposes. This pursuit is founded upon the willingness of participants to allow their genetic information to be pooled for analyses, but many have expressed concerns over the privacy of this genetic information. It has previously been shown that individuals can be identified from their anonymized genomic data and this has prompted researchers to look for additional security measures. Computer scientists Bonnier Berger and Sean Simmons have developed a new tool to help achieve this goal by using an approach called differential privacy. To increase privacy, a small amount of noise, or random variation, is added to the results of a user’s database query. Although the information returned would provide useful results, it would make it more difficult to conclusively connect this data to a patient’s identity. A similar method has been used by the US Census Bureau and the US Department of Labor for many years.
However, some scientists, including Yaniv Erlich, have concerns that adding noise to the dataset will reduce users ability to generate useful results. Erlich stated that “It’s nice on paper. But from a practical perspective I’m not sure that it can be used”. In the search for privacy, free form access to the data is limited. This “privacy budget” limits the number of questions that can be asked and excludes hundreds or thousands of locations in a genome. Additionally, because noise naturally increases error, it weakens the overall conclusion that can be drawn from the query. Simmons expects that answers will be close enough to be useful for a few targeted questions. The tradeoff for increased security is that databases protected this way could be instantly accessible and searchable, which cuts down on getting access to databases such as those managed by the National Institutes of Health. Simmons added that this method is “meant to get access to data sets that you might not have access to otherwise”. The group plans to continue to refine this method to balance the needs of researchers for access to these data sets while maintaining patient privacy. (Anna Nowogrodzki, Nature)
Have an interesting science policy link? Share it in the comments!